Nutrient Deficiency of Paddy Leaf Classification using Hybrid Convolutional Neural Network

For billions of people worldwide, enhancing the quantity and quality of paddy production stands as an essential goal. Rice, being a primary grain consumed in Asia, demands efficient farming techniques to ensure both sufficient yields and high-quality crops. Detecting diseases in rice crops is crucial to prevent financial losses and maintain food quality. Traditional methods in the agricultural industry often fall short in accurately identifying and addressing these issues. However, leveraging artificial intelligence (AI) offers a promising avenue due to its superior accuracy and speed in evaluation. Nutrient deficiencies significantly impact paddy growth, causing issues like insufficient potassium, phosphorus, and nitrogen. Identifying these deficiencies in paddy leaves, especially during the mid-growth stage, poses a considerable challenge. In response to these obstacles, a novel approach is proposed in this study—a deep learning model. The methodology involves gathering input images from a Kaggle dataset, followed by image augmentation. Pre-processing the images involves using the Contrast Limited Adaptive Histogram Equalization (CLAHE) model, while the extraction of features utilizes the GLCM model. Subsequently, a hybrid convolutional neural network (HCNN) is employed to classify nutrient-deficient paddy leaves. The simulation is conducted on the MATLAB platform, and various statistical metrics are employed to assess overall performance. The results demonstrate the superiority of the proposed HCNN model, achieving an accuracy of 97.5%, sensitivity of 96%, and specificity of 98.2%. These outcomes surpass the efficacy of existing methods, showcasing the potential of this AI-driven approach in revolutionizing disease detection and nutrient deficiency identification in paddy farming.